Adam Wulkiewicz

@barendgehrels @AlastairHolmes Well, we use larger types in some places but not all of them, and the library is not tested for this use case. So your suggestion to use...

Right, in general we tend to support non-true references so it's possible that this should be `typename boost::range_reference::type`. I have to check it to be sure. It's possible that real...

I skimmed through the code using these iterators and there is at least one place in the library that relies on the return of true reference. It's in the implementation...

@peter-popov Could you share your polygon class or at least something similar?

@peter-popov Sorry I missed that, thanks. I was curious why doesn't it compile for you so I made this fix: https://github.com/boostorg/geometry/compare/develop...awulkiew:fix/remove_point_iterator It seems to work properly for `distance(ls, poly)`, `covered_by(ls,...

In principle both points on pole should be treated as being equal so duplicated and one of them should be ignored AFAIR. So this tile should be treated as polygon...

Thanks for the info. I'll look into it next week.

@remster AFAIK the algorithm is Boost.Geometry-specific algorithm based on graph traversal. @barendgehrels was it taken from some paper or designed from scratch? I can reproduce the bug and currently am...

Thanks. I'll look into this as well. It looks like another issue. The result should be the US. But that's interesting. Just to be sure we look at this problem...

I'm talking about horizontal edges. When represented in Polygon they are not going along parallels so lines of the same latitude (https://en.wikipedia.org/wiki/Parallel_(latitude)). Instead they are going along great circles (https://en.wikipedia.org/wiki/Great_circle)....